Re: Commit message / hash in commitfest page.

Hi,

Is it possible to have commit-message or at least git hash in
commitfest. It will be very easy to track commit against commitfest
item.

--
Ibrar Ahmed

On 2019-04-11 11:36, Ibrar Ahmed wrote:
> Hi,
>
> Is it possible to have commit-message or at least git hash in
> commitfest. It will be very easy to track commit against commitfest
> item.
>

Commitfest items always point to discussion threads. These threads often
end with a message that says that the patch is pushed. IMHO, that
message would be the place to include the commithash. It would also be
easily findable via the commitfest application.

Erik Rijkers

> --
> Ibrar Ahmed

On Thu, Apr 11, 2019 at 2:44 PM Erikjan Rijkers <er(at)xs4all(dot)nl> wrote:
>
> On 2019-04-11 11:36, Ibrar Ahmed wrote:
> > Hi,
> >
> > Is it possible to have commit-message or at least git hash in
> > commitfest. It will be very easy to track commit against commitfest
> > item.
> >
>
> Commitfest items always point to discussion threads. These threads often
> end with a message that says that the patch is pushed. IMHO, that
> message would be the place to include the commithash. It would also be
> easily findable via the commitfest application.
>

+1

> Erik Rijkers
>
>
> > --
> > Ibrar Ahmed

--
Ibrar Ahmed

On Thu, Apr 11, 2019 at 02:55:10PM +0500, Ibrar Ahmed wrote:
>On Thu, Apr 11, 2019 at 2:44 PM Erikjan Rijkers <er(at)xs4all(dot)nl> wrote:
>>
>> On 2019-04-11 11:36, Ibrar Ahmed wrote:
>> > Hi,
>> >
>> > Is it possible to have commit-message or at least git hash in
>> > commitfest. It will be very easy to track commit against commitfest
>> > item.
>> >
>>
>> Commitfest items always point to discussion threads. These threads often
>> end with a message that says that the patch is pushed. IMHO, that
>> message would be the place to include the commithash. It would also be
>> easily findable via the commitfest application.
>>
>
>+1
>

I think it might be useful to actually have that directly in the CF app,
not just in the thread. There would need to a way to enter multiple
hashes, because patches often have multiple pieces.

But maybe that'd be too much unnecessary burden. I don't remember when I
last needed this information. And I'd probably try searching in git log
first anyway.

regard

--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com> writes:
> On Thu, Apr 11, 2019 at 02:55:10PM +0500, Ibrar Ahmed wrote:
>> On Thu, Apr 11, 2019 at 2:44 PM Erikjan Rijkers <er(at)xs4all(dot)nl> wrote:
>>>> Is it possible to have commit-message or at least git hash in
>>>> commitfest. It will be very easy to track commit against commitfest
>>>> item.

>>> Commitfest items always point to discussion threads. These threads often
>>> end with a message that says that the patch is pushed. IMHO, that
>>> message would be the place to include the commithash. It would also be
>>> easily findable via the commitfest application.

> I think it might be useful to actually have that directly in the CF app,
> not just in the thread. There would need to a way to enter multiple
> hashes, because patches often have multiple pieces.

> But maybe that'd be too much unnecessary burden. I don't remember when I
> last needed this information. And I'd probably try searching in git log
> first anyway.

Yeah, I can't see committers bothering to do this. Including the
discussion thread link in the commit message is already pretty
significant hassle, and something not everybody remembers/bothers with.

But ... maybe it could be automated? A bot looking at the commit log
could probably suck out the thread links and try to match them up
to CF entries. Likely you could get about 90% right even without that,
just by matching the committer's name and the time of commit vs time
of CF entry closure.

regards, tom lane

On 04/13/19 15:56, Tomas Vondra wrote:
> I think it might be useful to actually have that directly in the CF app,
> not just in the thread. There would need to a way to enter multiple
> hashes, because patches often have multiple pieces.

The CF app already recognizes (some) attachments in the email thread
and makes them directly clickable from the CF entry page. Could it do
that with commit hashes, if found in the body of an email thread?
Gitweb does that pretty successfully with commits mentioned in
commit messages, and github does it automagically for text in issues
and so on.

Maybe it could even recognize phrases like "commit 01deadbeef closes
cf entry" and change the cf entry state, though that'd be gravy.

Regards,
-Chap

On Sat, Apr 13, 2019 at 04:27:56PM -0400, Tom Lane wrote:
> Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com> writes:
> > On Thu, Apr 11, 2019 at 02:55:10PM +0500, Ibrar Ahmed wrote:
> >> On Thu, Apr 11, 2019 at 2:44 PM Erikjan Rijkers <er(at)xs4all(dot)nl> wrote:
> >>>> Is it possible to have commit-message or at least git hash in
> >>>> commitfest. It will be very easy to track commit against commitfest
> >>>> item.
>
> >>> Commitfest items always point to discussion threads. These threads often
> >>> end with a message that says that the patch is pushed. IMHO, that
> >>> message would be the place to include the commithash. It would also be
> >>> easily findable via the commitfest application.
>
> > I think it might be useful to actually have that directly in the CF app,
> > not just in the thread. There would need to a way to enter multiple
> > hashes, because patches often have multiple pieces.
>
> > But maybe that'd be too much unnecessary burden. I don't remember when I
> > last needed this information. And I'd probably try searching in git log
> > first anyway.
>
> Yeah, I can't see committers bothering to do this. Including the
> discussion thread link in the commit message is already pretty
> significant hassle, and something not everybody remembers/bothers with.
>
> But ... maybe it could be automated? A bot looking at the commit log
> could probably suck out the thread links and try to match them up
> to CF entries. Likely you could get about 90% right even without that,
> just by matching the committer's name and the time of commit vs time
> of CF entry closure.

I've been getting a lot of lift out of the git_fdw (well, out of
caching it, as performance isn't great yet) for constructing the
PostgreSQL Weekly News section on things already committed.

About 3.5% of commits (as of last week) on master are within a minute
of each other, so grabbing a window two minutes wide would work even
if we didn't have the committer's name in hand, it's unlikely to
produce more than one result.

Best,
David.
--
David Fetter <david(at)fetter(dot)org> http://fetter.org/
Phone: +1 415 235 3778

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

On Sat, Apr 13, 2019 at 10:28 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com> writes:
> > On Thu, Apr 11, 2019 at 02:55:10PM +0500, Ibrar Ahmed wrote:
> >> On Thu, Apr 11, 2019 at 2:44 PM Erikjan Rijkers <er(at)xs4all(dot)nl> wrote:
> >>>> Is it possible to have commit-message or at least git hash in
> >>>> commitfest. It will be very easy to track commit against commitfest
> >>>> item.
>
> >>> Commitfest items always point to discussion threads. These threads
> often
> >>> end with a message that says that the patch is pushed. IMHO, that
> >>> message would be the place to include the commithash. It would also
> be
> >>> easily findable via the commitfest application.
>
> > I think it might be useful to actually have that directly in the CF app,
> > not just in the thread. There would need to a way to enter multiple
> > hashes, because patches often have multiple pieces.
>
> > But maybe that'd be too much unnecessary burden. I don't remember when I
> > last needed this information. And I'd probably try searching in git log
> > first anyway.
>
> Yeah, I can't see committers bothering to do this. Including the
> discussion thread link in the commit message is already pretty
> significant hassle, and something not everybody remembers/bothers with.
>
> But ... maybe it could be automated? A bot looking at the commit log
> could probably suck out the thread links and try to match them up
> to CF entries. Likely you could get about 90% right even without that,
> just by matching the committer's name and the time of commit vs time
> of CF entry closure.
>

Would you even need to match that? It would be easy enough to scan all git
commit messages for links to th earchives and populate any CF entry that
attaches to that same thread.

Of course, that would be async, so you'd end up closing the CF entry and
then have it populate with the git information a bit later (in the simple
case where there is just one commit and then it 's done).

Unless we want to go all the way and have said bot actualy close the CF
entry. But the question is, do we?

--
Magnus Hagander
Me: https://www.hagander.net/ <http://www.hagander.net/>
Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>

On 2019-04-16 08:47, Magnus Hagander wrote:
> Unless we want to go all the way and have said bot actualy close the CF
> entry. But the question is, do we?

I don't think so. There are too many special cases that would make this
unreliable, like one commit fest thread consisting of multiple patches.

--
Peter Eisentraut http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On Tue, Apr 16, 2019 at 8:55 AM Peter Eisentraut <
peter(dot)eisentraut(at)2ndquadrant(dot)com> wrote:

> On 2019-04-16 08:47, Magnus Hagander wrote:
> > Unless we want to go all the way and have said bot actualy close the CF
> > entry. But the question is, do we?
>
> I don't think so. There are too many special cases that would make this
> unreliable, like one commit fest thread consisting of multiple patches.
>

I definitely don't think we should close just because they show up. It
would also require a keyword somewhere to indicate that it should be
closed. Of course, it can still lead to weird results when the same thread
is attached to multiple CF entries etc. So I agree, I don't think we'd want
that. Which means we'd have the async/out-of-order issue.

--
Magnus Hagander
Me: https://www.hagander.net/ <http://www.hagander.net/>
Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>

Magnus Hagander <magnus(at)hagander(dot)net> writes:
> On Tue, Apr 16, 2019 at 8:55 AM Peter Eisentraut <
> peter(dot)eisentraut(at)2ndquadrant(dot)com> wrote:
>> On 2019-04-16 08:47, Magnus Hagander wrote:
>>> Unless we want to go all the way and have said bot actualy close the CF
>>> entry. But the question is, do we?

>> I don't think so. There are too many special cases that would make this
>> unreliable, like one commit fest thread consisting of multiple patches.

> I definitely don't think we should close just because they show up.

Agreed.

> ... Which means we'd have the async/out-of-order issue.

I don't see that as much of a problem. The use-case for these links,
as I understand it, is for retrospective examination of CF data anyway.
The mere fact of closing the CF entry is enough for real-time status.

regards, tom lane